detecting declared character encoding of an XML file during parsing?

[bugbear]

I'm (successfully) parsing an XML file using SAX.
All the internal data passed to the various handlers
is (of course) in nice Java String format AKA
Unicode.

No problem.

For reasons of partability and/or courtesy
I would like to construct my "reply" XML file
in the same encoding as the "request" XML file
I just parsed.

So my question is: is there anyway from a SAX (or DOM?)
parser to find out what encoding was declared
in the XML file?

BugBear